Contrary to common misconceptions, American Sign Language (ASL) isn’t just signed English. It’s a completely distinct language with its own unique grammar system. This can make communication between those who know ASL and those who don’t challenging. Luckily, researchers at Cornell University are working to mitigate these conversational hurdles through wearable technology.
As they lay out in their research paper, they’ve developed the SpellRing—a quarter-sized device with a built-in microphone and speaker that’s meant to be worn on the thumb. Here’s how it works: The microphone and speaker release and pick up silent sound waves as the wearer moves their hands and fingers. Meanwhile, a tiny gyroscope also tracks hand motions. A deep-learning algorithm processes these sonar images and translates them into ASL letters in real time.
This isn’t the first technology designed to interpret finger gestures into sign language. However, existing devices tend to be bulky, making them difficult to use. Paper co-author Hyunchul Lim and other researchers at the Cornell Ann S. Bowers College of Computing and Information Science were inspired to look for solutions better suited for daily life.
According to the Cornell Chronicle, the SpellRing has similar accuracy compared to larger translation devices. The developers tested the gadget using 20 experienced and beginner ASL signers who were asked to spell over 20,000 words. The SpellRing’s accuracy ranged between 82 and 92 percent, varying based on the length and difficulty of the terms. Though not perfect, the numbers are still impressive. Training an AI to recognize 26 hand shapes for every letter of the alphabet is challenging, as signers naturally adjust their formation for the sake of flow.
To bring the SpellRing to the next level, Lim hopes to one day pair it with eyeglasses that process the upper body movements and facial expressions of signers. “Deaf and hard-of-hearing people use more than their hands for ASL,” he told the Cornell Chronicle. “ASL is a very complicated, complex visual language.”
Technology isn’t just good for helping people who already know ASL communicate—it can also be a learning tool for those new to the language. Nvidia’s free AI-powered tool, Signs, makes learning sign language easier by providing instant feedback during practice.
Read More About Language: